AI in Recruitment. Revolutionary or Rash?

Published: 8th May 2024 | Author: Kate Howes

Using AI tools in recruitment is a real ‘thing’ at the moment, and we imagine it will be for the foreseeable future.

For better or worse, though? Our gut instinct has always been to be wary of blindly using AI hiring tools – but then again, Gemma and I are pretty old school in our approach to how we carry out our services at Horizon. Our experience has been limited to hearing candidates (rightly) complain to us about heavily automated, digital hiring processes, and to our frustration at the volume of approaches we get from companies trying to sell us AI hiring tools to ‘revolutionise’ our recruitment agency. Realising we didn’t actually know enough about how AI can be used in recruitment, we wanted to explore its positives and its pitfalls…

If you boil it right, right down, many hiring processes are just that; processes to determine a match. Matching the skills one person (the candidate) has got to the requirements of another person (the hiring manager). We are matching candidate availability to client availability when arranging interviews, and those in lots of ways are the two biggest functions in hiring. Assessing suitability and availability – both processes that could, arguably be done by algorithms intelligent enough to mimic human behaviour.

More and more recruiters are turning to AI to outsource this part of their work; particularly CV searching and reviewing and initial ‘interviews’ to assess suitability. In our office, we are contacted several times a week by companies offering software that will ‘save us time and money’ by ‘taking the hassle out’ of these parts of our jobs. We’ve heard from candidates, as well, that applying for jobs and finding out about opportunities is an increasingly digital procedure, with often little to no human engagement until, if you make it through, you are finally invited to discuss the opportunity with a real person. Their frustrations are understandable, given that hiring is a two-way street, where communication between both parties should be straightforward. As much as you, the hirer, want to assess somebody’s suitability, surely you want any prospective employee to be able to assess whether they’d genuinely be happy doing the job that you’re hiring for?

While we are reluctant to shun advances in our sector and agree in many ways AI tools could actually benefit some parts of the hiring process – we don’t believe it should be relied on, and certainly not hailed as something that should revolutionize our sector, though it’s very likely it will become more commonplace in the coming years. The recruitment industry is centred around people and to remove the human element from our work seems utterly wrong to us.

While looking at sources to try and discover some ‘pros’ to offloading some of our workload to AI; there seems to be a focus on the fact traditional recruitment methods are labour-intensive and time-consuming. This is true, but we believe it kind of needs to be in many ways. We don’t want to save half an hour on asking AI text generators to write our client’s job adverts – not just because they’re dry and lack context – but because it’s important that we write decent, informative, detailed enough adverts so we catch the attention of the right candidates.

To us, it’s really important that we take the time to write good adverts, respond to people’s applications ourselves, are available at the end of a phone, and read every CV we come across. It’s literally what we get paid to do by our clients, and it’s respectful to our candidates.

It is said that using AI to review applications and filter them according to who is a skills match and who isn’t stops any (conscious or not) bias from the recruiter/hirer. This is an interesting concept, and one that could benefit our industry and recruitment practice as a whole. Digging a little deeper, it appears that AI isn’t quite there yet and in fact, in a study by Amazon*, their AI selection tool did in fact filter out women, and those without western-sounding names. After all, AI can only work with the data it is presented and can only mimic human behaviour. So, perhaps if we want to utilise this tool, we need to know how to feed it properly for it to work for us.

We have banged on, over the years about ‘good recruitment’ – it’s something we preach and practice. Good recruitment isn’t just about being efficient and effective at finding people for jobs, and professional but also about a personal touch – something AI tools can’t emulate. An important part of our work is empathising when someone calls up stunned by a shock redundancy or is navigating finding their first job after leaving education. It’s metaphorically ‘holding the hand’ of your candidate who is attending an interview for the first time in 20 years because they haven’t had to in the past. It’s commiserating with someone who loses out on an opportunity they desperately wanted, and celebrating with someone when they are offered the job they didn’t think they’d ever land.

So, while AI no doubt has its uses as a tool in our industry, let’s not get carried away. People are our industry, and to shirk away from working with them is to no longer be doing the job properly.